Reference · Videos — guarded, asynchronous text-to-video. You get a quote, submit a job, poll it, then download the file.

Overview

How a video job works

GET /v1/videos/models lists the models you can use now, each with a from_price (the cheapest clip it can make).
POST /v1/videos/quote prices your exact request without holding anything.
POST /v1/videos returns 202 with a job. The quoted price is held on your prepaid balance and charged only when the video is delivered. A failed, blocked or expired job is refunded.
Poll GET /v1/videos/{id} until status is completed, then GET /v1/videos/{id}/content redirects (302) to a short-lived download URL.

The price is the provider's price for that clip plus our margin, in US dollars (quote_usd) and, when a rate is available, in Rupiah (quote_idr). Signup credit and passes can't be used for video; it needs topped-up balance.

Free draft: every account with a verified email can make one free draft while the monthly free-draft budget lasts ("free_draft": true: Veo 3.1 Lite, 4 seconds, 720p, no audio).

Request

Body

curl https://api.anoman.io/v1/videos \
  -H "Authorization: Bearer anm-sk-..." \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: 6f1c2b9e-launch-001" \
  -d '{
    "model": "veo-3.1-lite",
    "prompt": "A calm lake at sunrise, slow aerial push-in",
    "seconds": 4,
    "resolution": "720p",
    "aspect_ratio": "16:9",
    "audio": false
  }'

Field	Description
`model`	A model id from `GET /v1/videos/models`, or `auto` (default) to let the gateway pick by `quality`.
`prompt`	The text prompt, up to 2,000 characters. Required.
`seconds`	Clip length. Each model lists the durations it supports. Required.
`resolution`	e.g. `720p` (default) or `1080p`, as the model lists.
`aspect_ratio`	`16:9` (default), `9:16` or `1:1`, as the model lists.
`audio`	Ask for an audio track where the model makes it optional.
`quality`	For `auto`: `draft`, `standard` (default) or `high`.
`negative_prompt`, `seed`	Optional, where the model supports them.
`max_price_usd` / `max_price_idr`	Refuse (`409 price_changed`) if the price is now above what you showed your user.
`free_draft`	`true` for the one free draft (overrides model, length, resolution and audio).

Send an Idempotency-Key header so a retry returns the same job instead of a second hold. Image-to-video is not available yet.

Limits & guards

Prompt guard. Child-safety, violent extremism, NSFW and content moderation (English + Bahasa Indonesia) run before anything is held; validated personal data is masked before the prompt leaves.
Output check. Frames of every result are checked for nudity before delivery; a blocked result is refunded.
Labels. Files are named …-ai-generated.mp4 and carry a C2PA manifest signed by Anoman (self-signed).
Where it runs. Generation runs at fal.ai, outside Indonesia (fal does not publish the region), under the model owner's terms (terms on each model).
Caps. A daily spend cap and an in-flight cap per workspace; a busy queue returns 503 and nothing is held.

Errors

Status	`error`	Meaning
402	`insufficient_balance`	Not enough topped-up balance for this quote.
409	`price_changed`	The price is above your `max_price_*`.
409	`free_draft_unavailable`	The free draft was already used or isn't available.
429	`video_daily_cap_reached` / `concurrent_limit`	A workspace cap is reached.
503	`capacity` / `video_temporarily_unavailable`	Try again later; nothing was charged.

Response

Illustrative example — the amounts are not a price list; the live quote comes from the API (POST /v1/videos/quote).

{
  "id": "vid_1a2b3c4d5e6f7a8b9c0d1e2f",
  "object": "video",
  "status": "queued",
  "model": "veo-3.1-lite",
  "seconds": 4,
  "_anoman": {
    "quote_usd": "0.1500",
    "quote_idr": 2775,
    "held": true,
    "poll_url": "/v1/videos/vid_1a2b3c4d5e6f7a8b9c0d1e2f",
    "label": "AI-generated"
  }
}

POST /v1/videos